13 June 2024 in IEEE Sensors Journal
The wheelset is a critical component of a rail vehicle, carrying the weight of the train in contact with the rail during the train operation. Constant measurement of wheelset geometries is required to ensure safe train operation. This paper presents a wheelset geometries measurement system based on multi-line structured light sensors. First, the principle of the system is introduced. Second, a two-dimensional calibration target with circular markers is designed to calibrate the camera and extract the centerline of the light bar in light interception image. A special bright block region is used to complete the structured light matching to complete the calibration of the light plane. Then, a method for obtaining the tread curve based on the radius of the circular cross-section is described to obtain wheelset geometries, converting the problem of fitting a three-dimensional sphere radius to a two-dimensional plane. The system was implemented at Shihu Station on Guangzhou Metro Line 14. The results show that the presented system can measure different wheelset geometries with an accuracy of 0.5 mm for wheel diameter and 0.2 mm for flange thickness and flange height, better repeatability and cheaper compared to other systems and manual measurements.
27 May 2022 in IEEE Transactions on Instrumentation and Measurement
The pantograph is one of the most critical electrical equipment in the electric locomotive as it transmits power from the contact wire to the locomotive. Condition monitoring of the pantograph is important to ensure the safe and consistent operation of a train. However, condition monitoring of the pantograph is difficult due to the complex lighting situation and background environment around the pantograph. In this article, we propose an image-based online monitoring system for pantograph wear and attitude. First, the working principle of the system is introduced. Second, an evaluation method of sliding plate wear state is presented. The wear curves are extracted based on image processing algorithms such as the sliding plate positioning and edge detection. Four criteria are calculated from the wear curves to evaluate four wear types. Then, a pantograph head attitude monitoring method based on pantograph horn boundary point extraction is introduced. After the pantograph horn detection, the horn boundary points are extracted to calculate the attitude parameters and then evaluate whether the pantograph head attitude is normal. The system was implemented in the Feieling metro station, Guangzhou Metro Line 9, China, and system performance was analyzed using real-time pantograph images. The experimental results have shown that the proposed system can detect the four wear types of the sliding plate and monitor the attitude of the pantograph head with a higher accuracy and better repeatability compared to other systems and manual measurements.
2 April 2020 in Lecture Notes in Electrical Engineering